Overview
What this challenge is about.
You receive a 12-month anonymized dataset of subscriber events (logins, lesson completions, payment history, support tickets) for around 200,000 users. Define churn precisely (no payment renewal within 14 days of period end), engineer features at the user-week grain, and train at least three model families (logistic regression, gradient-boosted trees, a small MLP). Pick a primary metric the business actually cares about (recall at top-decile lift, not raw AUC) and beat a naive recency baseline by at least 25%. Wrap the pipeline in a notebook + reproducible script set, and write a 2-page memo for the analytics lead explaining what the model can and cannot do.
The Brief
What you'll do, and what you'll demonstrate.
Deliver a reproducible, honestly-evaluated churn-prediction mini-project that beats the recency baseline on a business-aligned metric.
Earning criteria — what you'll demonstrate
- Scope a real ML problem from a vague business ask
- Implement and evaluate multiple model families fairly
- Pick metrics aligned with the downstream business action
- Document an ML mini-project so non-ML teammates can rerun it
Program Fit
Where this fits in your program.
Sharpens the same skills your degree expects you to demonstrate.
Skills
Skills you'll demonstrate.
Each one shows up on your verified credential.
Careers
Roles this prepares you for.
Real titles. Real skill bridges. Pick the one closest to your trajectory.
Machine Learning Engineer
Owning a churn project from problem framing to a reproducible pipeline that another teammate can rerun is the day-one work expected of a junior MLE on a small data team.
This challenge sharpens
- feature-engineering
- model-evaluation
- python
Data Scientist
Picking business-aligned metrics, calibrating probabilities, and writing the memo that explains what the model can and cannot do is the heart of applied data-scientist work.
This challenge sharpens
- model-evaluation
- data-cleaning
- feature-engineering
Applied AI Scientist
Comparing three model families on a real dataset and defending the winner in writing mirrors the applied AI scientist's job of mapping research methods onto product problems.
This challenge sharpens
- gradient-boosting
- pytorch
- model-evaluation